PPT Institute for Ethics and Emerging Technologies

Document Sample
PPT Institute for Ethics and Emerging Technologies Powered By Docstoc
					     Taxes on the Longevity Dividend:
           Can we Reduce Them?
Lessons from the Theoretical Foundations of
   Medical Cost-Effectiveness Analysis

             David Meltzer MD, PhD
              University of Chicago
• Improvements in health and resulting gains in
  quality of life and and longevity are highly valuable
• Value is offset to some degree by costs of longevity
• Magnitude of this offset in terms of welfare
  depends on size of gains in longevity compared to
  gains in quality of life
• Relative size of gains in length and quality of life
  will be affected by choices that we make
• Best decisions about health care resource allocation
  will reflect the relative costs of improving the
  length and quality of life
   Improved Health Highly Valuable
• Cutler and Richardson (1997), Murphy and Topel
  (2003), Nordhaus (2003)
• Multiply value of life-year saved (LYS) from
  statistical value of life (SVoL) based on revealed
  preference studies * increase in life expectancy from
  1970 to 2000
• Value D LY= 0.2 LYS (2-3months)/person/yr *
  $50,000/LYS = $10,000/person/yr
   – Comparable to increase in per capita GDP over entire
• $3 trillion per year, or $90 trillion from 1970-2000
   – Value D LY = 300 million people * $50,000/LYS * 6 LYS
• Estimates may be too low
   – Some estimates of SVoL ~$200,000/yr
   – Value of improvements in quality of life adds to this
     Value of Increased Longevity
     Offset by Costs of Longevity
• Benefits of improved health
   – Increased length and quality of life
   – Increased QALYs = D btSt Qt
      • St survival probability
      • Qt quality of life adjustment
      • b < 1 time preference discount factor
   – Increased years of productivity
   – Improved productivity and reduced health care costs at
     all ages
• Costs of improved health
   – Increased years of consumption and medical care
• Value of improved health = benefits - costs
 Consumption, Medical Expenditure,
Earnings and Net Resource Use by Age
Cost Offset of Health Gain Depends on
Gains in Longevity vs. Quality of Life
 • Gain LY = DQALY – DLY (C + M - E)
   – Cost offset = C + M - E
 • Gain QOL = DQALY
   – Cost offset = 0
 • Cost Offset per Gain in Health (DQALY)
   = (C + M - E) * (DLY/DQALY)
   = <$20,000 * (1/QOL)
      • < $50,000/QALY if QOL<0.4
      • < $100,000/QALY if QOL<0.2
Methods for Quality of Life Adjustment
  • Linear analog scale
  • Standard gamble
  • Time trade-off
    Linear Analog Scale

0                         1
Standard Gamble

             p       1

      Q=p            0
    Time Trade-off

              t      T
Diabetes-related Complication Utilities
        Complication        Mean (95% Cl)
           Angina           0.65 (0.63, 0.67)
        Mild Stroke         0.71 (0.68, 0.73)
        Severe Stroke       0.32 (0.29, 0.34)
    Peripheral Neuropathy   0.67 (0.65, 0.70)
        Amputation          0.56 (0.53, 0.59)
    Diabetic Retinopathy    0.54 (0.51, 0.57)
         Blindness          0.40 (0.37, 0.42)
    Mild Kidney Disease     0.66 (0.63, 0.69)
       Kidney Failure       0.36 (0.34, 0.39)
• Cost-offsets from increased longevity
  unlikely to cause gains in longevity to be
• But does this mean potential costs of
  longevity are irrelevant?
Cost-Effectiveness of Medical
Background: Accounting for Future Costs
• Save patient with medical care today who requires
  care in the future. Should we count that as a cost?
   – Related illness?
      • Angioplasty today, count bypass in future?
   – Unrelated illness?
      • Influenza vaccine today, count dialysis in future?
   – Non-medical costs and benefits?
      • Suicide prevention today, earnings in future?
        Consumption in future?
        Traditional Treatment of Future
          Costs and Benefits in CEA
• Analyses generally include:
  – Future benefits
      • Length of life/Quality of life = QALYs
  – Future medical costs for related illnesses
• Analyses generally exclude:
  – Future medical costs for unrelated illnesses
      • Few exceptions: Weinstein, OTA
   – Future non-medical costs
• Controversies reflect weak theoretical foundation of CEA
Theoretical Background: Phelps & Garber (1997)

  • Use lifetime utility maximization model
  • Conclude: Obtain same relative rankings of
    interventions if you include or exclude future
    medical costs for unrelated illness as long as they
        • treated consistently
        • truly unrelated = “conditional independence”
Theoretical Background: Meltzer (1997)
• Use lifetime utility maximization model
• Conclude:
   – Must include all future net resource use
      • Medical costs - both related and unrelated - and future non-medical
        costs net of earnings
      • “Net resource use”= consumption + medical expenditures - earnings
      • From -$10,000/ year @ age 25 to +$20,000/year @ age 85
   – Relative rankings of interventions not independent of
     future costs
      • Analyses that omit future costs favor interventions that extend
        life over those that improve quality of life
   – Phelps/Garber inadvertently assume net annual
     resource use is zero
• Consider two interventions with equal current cost
  that both produce one QALY
      • A increases life expectancy by one year at QOL=1
      • B increases life expectancy by two years at QOL=0.5
• Which is preferred?
   – From utility side: Indifferent
   – From cost side: A preferred since it saves the costs of
     supporting an extra year of life
   – Hence, A preferred overall
• Omitting future costs favors interventions that
  extend life (B) versus those that increase QOL (A)
Accounting for Future Costs
 Consumption, Medical Expenditure,
Earnings and Net Resource Use by Age
 Present Value of Future Net Resource
Use Per Year of Life Saved by Averting
            Death (by Age)
Approximate Effects of Future Costs
   Effects of Future Costs for
Interventions among the Elderly
    Implications for the Longevity Dividend
•   Biggest dividend is health itself
•   Longevity produces cost offset: C + M - E
•   Less important than health gains unless QOL very low
•   But cost-offset exists and can change cost-effectiveness
    of medical interventions
    – Recognition of this causes us to favor interventions that
      improve quality versus length of life
    – Reduces this “tax” on the longevity dividend
• Longevity dividend influenced by the choices we make

Shared By: